Stair climbing task of humanoid robot by phase composition and phase sequence

A goal of this work is to propose an algorithm that generates extended task of humanoid robot to endure a large variety of performance. In this approach, it is likely that using "task" library, which is constructed by previous task is effective method to generate extended task of humanoid robot according to environment change. It could be realized by method adopting "phase" transition, defined as "phase sequence". Also, it would realize to plan a task that required for humanoid robot in environment change using re-composition and re-using of constructed phase without motion experiments and planning from scratch. We have been applied a previous phase of step ascending and walking task based on strategy of actual motion for humanoid robot task, stair climbing. In the connection of two tasks, we adopted third-order bezier curve, which can be transformed desired trajectory using modifying its control parameters. Furthermore, appropriate task transition operation for phase sequence method was applied by re-composition of phases as variation of designated environments. As a result, extended task, stair climbing, was realized by recomposing of two fundamental tasks. In conclusion, a proposed task generation algorithm was verified.